3

Any set of four identical triangles can be arranged to form a four-surface solid. If these four triangles are not equilateral, e.g. 3,4,5, how do I find the altitude of from the one used for the bottom to the apex formed by the other three.

Blue
  • 83,939
poetasis
  • 6,795

5 Answers5

4

Since Tetrahedron is usually meant to designate the regular polyhedron, we can better say that we are constructing a triangular Pyramid.

Intuitively speaking, we are starting with two identical sets of three "poles", having different lengths but respecting the triangular law.

Piram_Triang_1

With one set we construct the base, $\triangle{ABC}$, and we designate, as usual, with $a$ the length of the edge opposed to $A$, with $\alpha$ the angle in $A$, etc.

Thereafter, with the second set of poles we are going to construct a "tent" over the base.
To do that, we order the poles in a non-decreasing order, e.g. $a,b,c$.
We place the shortest $a$ in $A$, the longest $c$ in $C$, and $b$ in $B$, and join at the top $V$.
We have obtained four exact copies of the triangle $ABC$, arranged congruently.
And there is no other way to arrange them.
The sum of the angles in $V$ is $\pi < 2\pi$, so everything looks to be in place.

Now we place a reference system with the origin in one vertex, and one edge of the base along the $x$ axis, as in the sketch.
To obtain the position of $V$, we impose to the vector $\bf b = \vec{BV}$ to make an angle $\gamma$ with $\bf a = \vec{BC}$ and an angle $\alpha$ with $\bf c = \vec{BA}$, and to have a length of $b$.
That means $$ \left\{ \matrix{ \left( {\matrix{ a & 0 & 0 \cr {c\cos \beta } & {c\sin \beta } & 0 \cr } } \right) \left( {\matrix{ {v_x } \cr {v_y } \cr {v_z } \cr } } \right) = \left( {\matrix{ {ab\cos \gamma } \cr {cb\cos \alpha } \cr } } \right) \hfill \cr \left( {\matrix{ {v_x } & {v_y } & {v_z } \cr } } \right) \left( {\matrix{ {v_x } \cr {v_y } \cr {v_z } \cr } } \right) = b^{\,2} \hfill \cr} \right. $$ which is readily solved as $$ \bbox[lightyellow] { \left\{ \matrix{ v_x = b\cos \gamma \hfill \cr v_y = b\left( {\cos \alpha - \cos \beta \cos \gamma } \right)/\sin \beta \hfill \cr v_z = b\sqrt {1 - \cos ^{\,2} \gamma - \left( {\cos \alpha - \cos \beta \cos \gamma } \right)^{\,2} /\sin ^{\,2} \beta } \hfill \cr} \right. }\tag{1}$$

$v_z$ is the heigth to compute the volume, which thus will be $$ \bbox[lightyellow] { \eqalign{ & V = {1 \over 6}abc\sqrt {\left( {1 - \cos ^{\,2} \gamma } \right)\sin ^{\,2} \beta - \left( {\cos \alpha - \cos \beta \cos \gamma } \right)^{\,2} } = \cr & = {1 \over 6}abc\sqrt {1 + 2\cos \alpha \cos \beta \cos \gamma - \left( {\cos ^{\,2} \gamma + \cos ^{\,2} \beta + \cos ^{\,2} \alpha } \right)} \cr} }\tag{2}$$ which can be then reformulated in various ways by using the sine or cosine laws.

But .., whoever has practically built any such a tent, or origami, or else, might smell that it can't go always so .. straight.
And in fact, if one of the angles, e.g. $\gamma$, is rect, in the above we get $v_x=0$.
Some investigation shows that the tent has become flat, with $AVBC$ that becomes a rectangle.
And some more investigation shows that the construction (always dealing with four equal triangles) is:
- possible when all the angles are less than $\pi/2$ (acute triangles),
- becomes flat if one angle is $\pi/2$,
- and is impossible if the triangles are obtuse.

Note that, by replacing $\alpha$ with $\pi - \alpha -\beta$ and then putting it back, we can rewrite the result above as

$$ \bbox[lightyellow] { \eqalign{ & V^{\,2} = {1 \over {36}}a^{\,2} b^{\,2} c^{\,2} \left( {\left( {1 - \cos ^{\,2} \gamma } \right)\sin ^{\,2} \beta - \left( {\cos \alpha - \cos \beta \cos \gamma } \right)^{\,2} } \right) = \cr & = {1 \over {36}}a^{\,2} b^{\,2} c^{\,2} \left( {\sin ^{\,2} \gamma \sin ^{\,2} \beta - \left( {\cos \left( {\pi - \left( {\beta + \gamma } \right)} \right) - \cos \beta \cos \gamma } \right)^{\,2} } \right) = \cr & = {1 \over {36}}a^{\,2} b^{\,2} c^{\,2} \left( {\sin ^{\,2} \gamma \sin ^{\,2} \beta - \left( {2\cos \beta \cos \gamma - \sin \beta \sin \gamma } \right)^{\,2} } \right) = \cr & = {1 \over {36}}a^{\,2} b^{\,2} c^{\,2} \left( { - 4\cos ^{\,2} \beta \cos ^{\,2} \gamma + 4\cos \beta \cos \gamma \sin \beta \sin \gamma } \right) = \cr & = {1 \over 9}a^{\,2} b^{\,2} c^{\,2} \cos \alpha \cos \beta \cos \gamma \cr} }\tag{2.a}$$ which coincides with the more mathematically elegant method provided by Blue.
And we can rewrite more compactly also the (1).

G Cab
  • 35,964
  • Excellent answer, I think. It will take me time to absorb it enough to apply it. Now, will it work for an open-base, three-sided pyramid made of three dissimilar triangles such as: (725,108,733) (333,644,725) (315,108,333)? – poetasis May 12 '18 at 13:38
  • @poetasis: well, the drawing shall help you enough. Then all you need is to know dot product (and trig of course). – G Cab May 12 '18 at 23:12
  • @poetasis: concerning your second question, we don't need dot product and trigonometry in this case. As far as you can lay the triangles on ground, arrange them (is flipping allowed ?) side by side (touching edges with same length), and end with two far edges of same length, then you can zip these together and get your tent. Pole-ends touching the ground are three, and there is always a plane that can sustain them. In your example: (725,733,108), (108,315,333),(333,644,725). – G Cab May 12 '18 at 23:14
  • Four of any identical acute triangle will make a three-side pyramid with a closed bottom but, programmatically, I have not found that to be the case for dissimilar triangles. Three-sided pyramids with open bottoms can be formed from dissimilar Pythagorean triplets and that is what I am exploring. – poetasis May 12 '18 at 23:53
  • @poetasis: "three-sided pyramids with open bottoms formed from dissimilar Pythagorean triplets" is an entirely different class of figures from tetrahedra formed from four congruent triangles. If you're exploring the former, then why ask us about the latter? In any case, the former can be aided by Menger's Theorem: Given six potential edge-lengths, if the Cayley-Menger determinants (linked in my answer) give positive face-area-squared and volume-squared values, then the lengths form a valid tetrahedron. You can get the altitude to the base of area $T$ via $3V=h T$. – Blue May 13 '18 at 02:18
  • @GCab: +1. However, I must object to your first sentence. "tetrahedron" is a perfectly acceptable term for the generic, not-necessarily-regular solid; for the regular version, we typically write "regular tetrahedron". (I know of what I write here. "hedronometry" ---the trigonometry of tetrahedra--- has been something of an obsession of mine for decades. :) – Blue May 13 '18 at 02:32
  • Perhaps I should have asked about the open bottom irregular three-faced pyramid first. I just thought it would be simpler to start with 3,4,5s instead of (725,733,108), (108,315,333),(333,644,725), etc. Worse still is the question of a three-sided open-bottom pyramid made of 6 dissimilar triangle such as / 205,133,156 | 156,667,685 \ / 685,37,684 | 684,1363,1525 \ / 1525,1517,156 | 156,133,205 . Wanna try to find the altitude on that one? – poetasis May 13 '18 at 10:00
  • @Blue: not being native english speaker, I take note of your objection and experience in hedronometry (did not know of this subject before). And in consideration of that, have you ever dealt with the Integral Volume of (irregular) Tetrahedra ? do you have any paper to suggest ? – G Cab May 13 '18 at 10:36
  • 1
    @poetasis: you have better to open another post specifically on this new subject, and put a link to it here. – G Cab May 13 '18 at 10:48
  • @GCab: I may well be the only person currently studying hedronometry. :) My work is more purely-geometric than number-theoretic, so I don't have any particular insight on (or references concerning) integral volume issues ... especially not if you mean in the context of tetrahedra that have integer side-lengths, integer face-areas, AND integer volume. That said, I do have formulas that relate face-areas and volume; see, for instance, my note "Heron-Like Hedronometric Results for Tetrahedral Volume". – Blue May 13 '18 at 11:30
  • 1
    @poetasis: The solution to your general three-sided-open-bottom problem is exactly as in my comment to your question: use the CM determinants to find volume and base-area, then use these to get altitude. That approach is completely general, applying when the triangles are congruent or dissimilar, right or not-right; and it's dead-simple. Your six-sided-open-bottom problem would take a bit more work; it's not obvious (or likely) that the six "bottom" edges lie in a common plane, so "altitude" may not even be defined. In any case, as GCab mentions, you should post these as separate questions. – Blue May 13 '18 at 11:47
  • @Blue: your note is really interesting, also because it is a non common subject. As for the subject of my interest (just amateurial) it concerns the integer volume, by somebody called "content", = number of integer points within tetrahedra with integer vertices (extension to 3D of Pick' theorem). – G Cab May 13 '18 at 12:51
2

Elaborating on my comment ...

Let the triangle have side-lengths $a$, $b$, $c$, and area $T$. The Cayley-Menger determinant for area ---equivalently Heron's formula--- tells us $$16 T^2 = (-a+b+c)(a-b+c)(a+b-c)(a+b+c) \tag{1}$$ The CM determinant for volume gives $$288V^2 = \left|\begin{matrix} 0 & 1 & 1 & 1 & 1 \\ 1 & 0 & a^2 & b^2 & c^2 \\ 1 & a^2 & 0 & c^2 & b^2 \\ 1 & b^2 & c^2 & 0 & a^2 \\ 1 & c^2 & b^2 & a^2 & 0 \end{matrix}\right| = 4 (-a^2 + b^2 + c^2) (a^2 - b^2 + c^2) (a^2 + b^2 - c^2) \tag{2}$$

Observe that, if the triangle has a right angle, then $(2)$ vanishes; the tetrahedron has zero volume. Moreover, if the triangle has an obtuse angle, then $(2)$ is negative, making $V$ imaginary: in this case, there must be no tetrahedron at all.

Since altitude $h$ satisfies $3 V = h T$,

$$h^2 = \frac{9 V^2}{T^2} = \frac{2(-a^2+b^2+c^2)(a^2-b^2+c^2)(a^2+b^2-c^2)}{(-a+b+c)(a-b+c)(a+b-c)(a+b+c)} \tag{$\star$}$$

For something more angle-related ... By the Law of Cosines, we have $$a^2 = b^2 + c^2 - 2 b c \cos A \qquad\text{etc}$$ so that $$V^2 = \frac{1}{72}\cdot 2 b c \cos A \cdot 2c a \cos B \cdot 2 a b \cos C = \frac{1}{9} a^2 b^2 c^2 \cos A \cos B \cos C$$ By the Law of Sines, $$a = d \sin A \qquad\text{etc}$$ where $d$ is the circumdiameter of the triangle, so that $$T = \frac12 a b \sin C = \frac12 d^2 \sin A \sin B \sin C \qquad V = \frac13 d^3\sin A \sin B \sin C\sqrt{\cos A \cos B \cos C} $$ Hence,

$$h = \frac{3V}{T} = 2d\sqrt{\cos A \cos B \cos C} \tag{$\star\star$}$$

Blue
  • 83,939
1

You have 3 legs for 1 triangle and it is repeated to make 4 triangles, which fit together to make some wierd tetrahedron. Therefore, we are given 3 of 4 total points. Assume them to be $(A_x,A_y,A_z),(B_x,B_y,B_z),(C_x,C_y,C_z)$. Once fitted, the fourth point, $D(x,y,z)$, can be found by some combination for a solution to the following three equations. $$\Vert (x,y,z)-A)\Vert = \Vert(B-A)\Vert$$ $$\Vert (x,y,z)-B\Vert = \Vert(C-A)\Vert$$ $$\Vert(x,y,z)-C\Vert=\Vert(B-C)\Vert$$ Using real 3d points for A,B,C Geogebra easily comes up with the two options for point $D$. From here it is just more analytic geometry. We are looking for the distance from the new point, $D$, to the plane made by points $A,B,C$.

The plane equations normal vector $(n_x,n_y,n_z)$ is given by getting the cross product of two vectors that are in the plane. The plane is then $$\left(\begin{array}{c}x\\y\\z\end{array}\right)\cdot\left(\begin{array}{c}n_{x}\\n_{y}\\n_{z}\end{array}\right)=\left(\begin{array}{c}A_{x}\\A_{y}\\A_{z}\end{array}\right)\cdot\left(\begin{array}{c}n_{x}\\n_{y}\\n_{z}\end{array}\right).$$ A line from point $D$ to the plane can be written as $$\left(\begin{array}{c}x\\y\\z\end{array}\right)=D+\lambda\left(\begin{array}{c}n_{x}\\n_{y}\\n_{x}\end{array}\right)$$

From here you can plug in $D$ and $n$ and get values for $(x,y,z)$ in terms of $\lambda$. i.e. little equations. Substitute these into the plane equation and solve for $\lambda$. Finally, plug lambda back into the line equation to get the intersection of the plane and the line, call it point $E$. $Distance = |D-E|$. It is a lot easier with numbers and probably easier to understand as well.

Narlin
  • 1,251
1

enter image description here

I had previously brought up the fact that I thought 4 identical triangles needed to be isosceles triangles to make a tetrahedron. I've added this sketch to show how isosceles triangles fold up to make a tetrahedron with the 3 apexes intersecting at a single point when folded up. Conversely, with a 3,4,5 triangle, when folded up, the apexes do not translate to a single point. Or am I missing something? Update, per the additional graphic, I guess it only works for acute triangles.

Phil H
  • 5,689
  • I want nothing to do with isosceles triangles even though ANY four identical triangles will form a three-sided closed-bottom pyramid. Here, we've been talking about Pythagorean triplets as in your middle sketch and, if you cut and fold a 4,6,8 out of cardboard, you will see that it makes a pyramid with 3,4,5 faces. – poetasis May 12 '18 at 23:45
  • how do you get 3,4,5 out of 4,6,8? – Phil H May 13 '18 at 01:09
  • @PhilH: When using a right triangle, the resulting tetrahedron is "flat": its edges coincide with the edges and diagonals of a rectangle. To reconcile that with your folding diagram, simply observe that the "upward" fold for that bottom triangle should create a dihedral angle of $180^\circ$ along the hypotenuse; that is, there's no fold at all. – Blue May 13 '18 at 01:55
  • Yes, and a straight line is a triangle whose 2 shorter sides sum to the longer side. Really :) – Phil H May 13 '18 at 02:03
  • I've replied in other places that I made a typo with 4,6,8; it should have been 6,8,10 and 6,8,10 does fold up into a 3-sided pyramid with a closed bottom, with all sides identical (3,4,5) but oriented differently. – poetasis May 13 '18 at 10:05
  • I am embarrass to say that 6,8,10 does not fold up into a 3,4,5 pyramid as I thought it would. My humble apologies to everyone. – poetasis May 13 '18 at 10:15
0

For 4 identical isosceles triangles, the height will be:

$$h = \sqrt{(b\cdot cos(A/2))^2-(b\cdot sin(A/2)tan(A/2))^2}$$

Where b is the base length, and A is the apex angle of the isosceles triangle.

Example: for 4 isosceles triangles with base 5 and apex angle 40 deg.....

$h = 4.657$

Phil H
  • 5,689
  • They are not isosceles triangles – poetasis May 12 '18 at 13:19
  • @poetasis In my example, the base is 5 the apex angle is 40 degrees and the base angles are 70 degrees. Why do you think they are not isosceles? – Phil H May 12 '18 at 13:29
  • None of the triangles have two equal sides. They are in fact Pythagorean triplets and the edges of the upper part of the pyramid are 3,4, and 5, respectively. – poetasis May 12 '18 at 14:25
  • People draw 3D wireframes of this shape but if they actually developed a flat pattern of 3,4,5 triangles and tried to fold up the triangles to make a tetrahedron....... well, don't take my word for it, give it a shot. – Phil H May 12 '18 at 16:42
  • cut a 4,6,8 triangle our of manilla paper and fold it up. You will see that the three sides around the 'base' to make an apex at a single point. – poetasis May 13 '18 at 00:02
  • yes but what about 3,4,5? – Phil H May 13 '18 at 01:05
  • I typed incorrectly. I meant that you should cut out a 6,8,10 triangle which is four 3,4,5s as in your middle drawing above. It should fold up to make the pyramid in question. – poetasis May 13 '18 at 01:35
  • 4,6,8 was a typo and should have been 6,8,10 – poetasis May 13 '18 at 01:38
  • 1
    4 identical right triangles don't work in creating a tetrahedron or any identical triangles with an obtuse angle. For 4 identical triangles, only acute triangles make a tetrahedron. Read G Cab's answer. – Phil H May 13 '18 at 01:47